A proof of the Riemann conjecture - Fuchs

The solution concept to answer the RH
Klaus Braun
Dec 20, 2014
www.riemann-hypothesis.de
The Riemann Hypothesis states that the non-trivial zeros of the Zeta function all have real part one-half. The
Hilbert-Polya conjecture states that the imaginary parts of the zeros of the Zeta function corresponds to
eigenvalues of an unbounded self-adjoint operator. In [BeM] some evidence is presented that the
eigenvalues are energy levels, that is eigenvalues of a hermitian quantum (“Riemann”) operator associated
with the classical Hamiltonian
Hcl ( x, p)  xp ,
where x is the (one-dimensional) position coordinate and y the conjugate momentum.
All attempts failed so far to represent the Riemann duality equation in the critical stripe as convergent Mellin
transforms of an underlying self-adjoint integral operator equation. The solution concept to answer the RH is
about the following:
1. to define two self-adjoint, bounded (singular integral) operators (which correspond to two convolution
integral representation [CaD]) with appropriate Hilbert space domains enabling proof P1 (Gaussian
function based) and proof P2 (fractional part function based)
2. to apply the special degenerated hypergeometric (Kummer) function

1 3
1 ( x) n 1 1 / 2  xt
1
F ( ; , x )  
  t e dt 
2 2
2
n

1
n
!
2
x
n 0
0
1
1 1
x
e
t 2
dt
0
for the definition of modified integral exponential and Gamma functions (in the critical stripe) by


E ( x) :   e  t d log t
1 3
E * ( x) :21 F1( ; , x)
2 2
x
s 
( )   x s / 2 dE
2 0


s
s s .
* ( ) :  x s / 2 dE *  ( )
2
2
s 1
0
For the Hilbert transform f H of the Gaussian function f it holds in the neighborhood of s  1
1 3 2
s
s
1
s
x 2
M  f H ( s)   (1 s ) / 2 tan( )( ) 
( ) with f H ( x)  4xe 1 F1 ( ; , x ) .
2 2
2 2
s 1 2
The functions f H and f are identical in a weak L2 (, ) ensuring appropriate relationship to the RH theory
([EdH]). The same correlation is valid for the fractional part function and its Hilbert transform


2
2
 ( x) : 2 ( x  x)     sin 2nx ,  H ( x) : 2 log 2 sin(x)   cos 2nx
1
n
1
n
with respect to the Hilbert space L#2 (0,1) ([TiE]).
The above enables appropriate integral operator representations of (s  1) (s) on the critical line and an
alternative Li-function ((EdH) 1.15) with appreciated approximation properties in the form

1 3
1
log n x
Li ( x) :21 F1 ( ; , log x)  
  xt t 1 / 2 dt
2 2
n!
0 n  1/ 2
0
1
1
.
Proof 1
The Poisson summation formula applied to the Gauss-Weierstrass density function
f ( x) : ex
2
enables the proof of the Jacobi’s   relation xG( x)  G(1 / x) for the Theta function

G ( x) : 1   ( x 2 ) :  f (nx) .

Let M denote the Mellin transform then it holds
M  f ( s ) 
1 s / 2 s .

( )
2
2
The natural statement of the functional equation is the symmetric duality equation ([EdH], 1.6)
 (s)M  f (s)   (1  s)M  f (1  s) .
(*)
Multiplying by s( s  1) and putting
1
2
s
2
 ( s) : s( s  1) ( s) ( )  s / 2
leads to the Riemann duality equation in the form
 (s)   (s)(s 1)M xf ( x)(s)   (1  s) , s  C .
Due to the corresponding property of the Hilbert transform the functions f , f H are identical in a
weak L2 (, )  sense, i.e. it holds
i)
f
2
0
 fH
2
0

1
2
ii)
( f ,  )  ( f H ,  )   L2 (, )
iii)
( f , f H )  0 , xH  Hx f H  0 .
From this it follows the identity
Fs : M  f (s)  M  f H (s) : FsH
in the sense of distributional complex-valued functions ([PeB] I.15).
Remark: It holds ([TiE] 2.14, [BeB] 6, example 1)
 ( s) 
1
1 
1
1
s  1
1

 s   

s 1
s  1  2 (n  1) s 1 n s 1
n  s 1
2
.
There is an only formally valid representation of Riemann’s duality equation as transform of an
integral operator in the form ([EdH] 10.3):
(**)  (s)  s  1 x1 sG( x) dx   x1 sG( x) dx  .

s  1 2 0

x

x
1
This operator has no transform at all, as the integral
1
x
0
1 s

dx
dx
1
1
 x1 s  formally , only (!) 

0
x 1
x
1 s 1 s
does not converge for any s. The integral (**) would converge if the constant term
f (0)  fˆ (0)  1
is absent. In order to overcome this issue to build series representation of  (s) as an even function
of s  1 / 2 , Riemann introduced the auxiliary function ([EdH] 10.3)
H ( x) :
d  2 d

x
G ( x)
dx  dx

which fulfills

2 ( s)   x1 s H ( x)
0
dx .
x
The Hilbert transform “spirits away” the jeopardizing constant Fourier series term ([PeB] Example
9.11)], i.e. the related Theta function G H has a vanishing constant term.
The Hilbert transform of the Gaussian function f (x) is given by ([GrI] 3.952, 7.612, 9.21-9.23)

1 3
f H ( x)  4  f ( ) sin(2x)d )  4xf ( x)1F1 ( ; , x 2 ) 
2 2
0
4
x
e
1
 x 2
2
M 1 1 (x 2 ) .
,
4 4
With respect to

2 xn
n  0 2n  1 n!
e  x F (  x)  e  x 
we refer to the following auxiliary function of ([BeB] III, Entry 10):

 ( x) : e x 
n 0
 ( n) x n
n
n!
which fulfills the Borel summable condition.
It holds


dx
s
s
M  f H ( s)   x f H ( x)
  (1 s ) / 2 tan( )( )   3 / 2
x
2
2
0
s
tan( s)
2 M xf ( x)( s)   3 / 2

2
s

1
M  f ( s)  c tan( (1  s))M  f H ( s)  M xf ( x)( s) .
2
s
3
s 0
The key idea is to replace

s
 x f ( x)
0


dx
1 s
dx
1 3
   s / 2 ( )   x s f H ( x)
 4  x s f ( x)1 F1 ( ; ; x 2 )dx
x
s 2
x
2 2
0
0
and
G ( x)  G H ( x )
to build a modified functional equation, being valid in a weak sense, but with same set of Zeta
function zeros.
With density arguments in combination with the Hardy theorem ([EdH] 11.1: the number of zeros of
the Zeta function on the critical line is infinite) the weak functional equation is then also valid in a
strong sense.
The prize to be paid is a RH analysis in a weak L2 (, )  Hilbert space framework and (weak)
variational representations of affected functions and dual integral operator equations. As this is
anyway the natural framework to deal with singular integral operators the prize seems to be more
than adequate.
Let
E ( x) : e x ,  (n) : 2n
2n  1
1 3
( x) : xF ( x) : 2 x1F1( ; , x) .
2 2
Lemma: It holds

E ( x )   ( x )    ( n)
i)
0
, Re ss   n (s)  (1)
( x) n
n!
n!
ii)
E (x) decreases faster than (x) for all x  0
iii)
Ei ( x) :   et

x
iv)
n

dt
 F ( x)    dF
t
x

Li( x)  Ei (log x)  Li  ( x) : F (log x)  
0
 (n) log n x .
n
n!
Lemma: ([GrI] 7.612) It holds
i)
ii)
M E (s)  (s)
M ( s)  ( s)
2s
 ( s) ( s)
2s  1
for
0  Re( s) 
1 ,
2
which corresponds to the Ramanujan Master Theorem representation ([BeB] IV,
Entry 11) resp.

1 3
s/2
0 x d 1F1 ( 2 ; 2 , x) 
s
s
(1  )  ( )
2 
2
s 1
s 1
for
0  Re( s)  1
,
which corresponds to the Gauss notation of the Gamma function ([EdH] 1.3).
4
Remark: From [SeA] we note that all zeros z n of

1 3
1 z n 1 zt 1 / 2
F ( ; , z)  
  e t dt  z 3 / 4e z / 2 M 1 1 ( z )
,
2 2
20
n  0 2n  1 n!
4 4
1
1 1
lie in the half-plane Re( z )  1 / 2 and in the horizontal stripe (2n  1)  Im( z )  2n . It holds
s
( )
1 3
dx
2
0 x 1 F1 ( 2 ; 2 , x) x  s  1

s/2
.
Remark: With respect to [PoG1] we note that
z

2
1 3
z
4zf ( z )1 F1 ( ; , z 2 )  czez  (1  )e n
2 2

1
n
is a function of the Laguerre-Polya class, i.e. a function of increased genus 1 (see also [CsG]).
Remark (Yukawa potential): The Ei ( x)  function is also related to the Yukawa potential of a
point charge in the form ([DuR])
e  r / r ,   0
in order to define a nuclear force potential which decays rapidly at infinity. Yukawa assumed that
 1 was of the order of magnitude of a nuclear radius. It results that the potential u of a charge
distribution satisfies the Yukawa equation
u   2 u
at points of free space. Thus the Yukawa potentials are invariant under the group of rotations and
translations of space, like those of Newton potentials. As  approaches zero the Yukawa
potentials approach those of Newton. The function e  r / r is a member of the Bessel family of
functions. Bessel functions have certain advantages over Newtonian potentials in functional
analysis (([DoW]).
Analog to the above we propose a modified Yukawa potential dY  by the substitution
e  r / r  F (r ) .
1
 the radial function
2
f (r ) of the corresponding eigenfunction of the Schrödinger equation is given by ([FlS] p 99)
Remark (ball symmetric potential of linear oscillator): for the energy E 
f (r )  e

r2
2
1 3
1
1 3

2
2 
c1 1 F1 ( 2 ; 2 , r )  c2 r 1 F1 ( 2 ; 2 , r )
5
Proof 2
The same idea of proof 1 can also be applied to the fractional part function in an alternative Hilbert
space framework ([KBr]):
Let H  L*2 () with  : S 1 ( R 2 ) , i.e.  is the boundary of the unit sphere. Let u (s) being a
2  periodic function and  denotes the integral from 0 to 2 in the Cauchy-sense. Then for
u  H : L2 () with  : S 1 ( R2 ) and for real  Fourier coefficients and norms are defined by
u :
1
u ( x)eix dx
2 
u
2


:  
2
u

2
.
Then the Fourier coefficients of the convolution operator
( Au )( x) :  log 2 sin
x y
u ( y)dy :  k ( x  y )u ( y)dy
2
and D( A)  H A  H 1 / 2 () .
are given by
( Au )  k u 
1
u .
2
The operator A enables characterization of the Hilbert spaces H 1 / 2 and H 1 in the form

H 1 / 2   
2
1 / 2


 ( A , ) 0   , H 1   
2
1

 ( A , A ) 0  
.
From ([BeB] 8, Entry 17 (iv)) we quote
“Ramanujan informs us to note that

cot(x)  2 sin(2 x)
,
1
which also is devoid of meaning, .... may be formally established by differentiating the equality

2
cos 2 x

1
 2 log 2 sin(x)
.
With respect to the Dirac function we note that building on the Dirichlet kernel there is a formal
representation of  (x) in the distribution sense in the form
 ( x) 
1
2

 einx 
n  
1
2
2
ikx
 e dk 
0
1

1
cos(kx)dk  sgn( x)  H

2
0
6
1 / 2 
( ,  )  H 1 ( ,  ) .
For

 ( x) : 2 ( x  x)     sin 2nx
1
2
n

2
n
 H ( x) : 2 log 2 sin(x)   cos 2nx
1

 : cot( )  2 sin(2n)
1
it holds (see also [ZyA] XIII, (11-3))

( A )( x)  


1
sin(2nx)
 n ( x)  
 H 0#
2n
n
1
.
From literature (e.g. [GaD] pp.63, [GrI] 1.441) we recall
1
2

0  2
1

1

e
in

e
0  2

0  2
e
in
 1 in
n  1,2,3,...
 2 n e

ln
d  
0
n0
 
 1 ein n  1,2,..
2 sin
2
 2n

in
1
 ie in n  1,2,3,...
1
 

cot
d   0
n0
2
2
 ie in n  1,2,..

 ne in

d   0
2  
 ne in
4 sin

2
1
n  1,2,3,...
n0
n  1,2,..
.
Due to the corresponding property of the Hilbert transform the functions  , H are identical in a weak
L#2 (0,1)  sense, i.e. it holds
i)

2
0
 H
2
0
ii)
( ,  )  ( H ,  )   L#2 (0,1)
iii)
( H ,  ) 1 / 2  ( ,  ) 1 / 2   L#2 (0,1)
because of ( H ,  ) 1 / 2  ( A H ,  ) 0  ( H ,  ) 0  ( A ,  ) 0  ( ,  ) 1 / 2 .
In the same way as for the P1 Hilbert space framework the prize to be paid is a RH analysis in a
weak H #1 / 2 (0,1)  Hilbert space framework ([PeB] I, §15, [BeJ]) and (weak) variational
representations of affected functions and dual integral operator equations (which is anyway the
natural framework to deal with singular integral operators and spectral theory):
In a weak H #1 / 2 (0,1)  sense it holds
 ( s)( s) cos(

2
s)  M  log 2 sin( )( s)   ( s) M cos ( s)
7
for 0  Re( s)  1 .
The relationship to the Riemann duality equation is given by ([TiE] 2.1)
2
M cos ( s) :  ( ) 1 s  (1  s)

with
 (1  s)  (s)  1
and
 (s)   (s) (1  s) .
The distributional P2 Hilbert scale framework is proposed to be applied to other areas, as well.
Based on the P2 Hilbert space framework in [Br1] an alternative framework for the HardyLittlewood circle method is given to analyze additive prime number problems.
Remark 1: There is a still open characterization question raised by Riemann ([RiB] p. 18) about
the representation of a function f(x) as convergent Fourier and/or trigonometric series ([ZyA] XII, 6,
7, 11):
“Wenn eine Function durch eine trigonometrische Reihe darstellbar ist, was
folgt daraus über ihren Gang, über die Aenderung ihres Werthes bei stetiger
Aenderung des Arguments?“
Remark 2: The dual space of H *1 / 2  H1 / 2  L2 is isometric to the classical Hardy space H2 of
analytical functions in the unit disc with norm
f (re i )
H2
:
1
2

 f (re
i
2
) d .

It holds
i) If f  H2 , then there exists boundary values f (e i )  lim f (rei )  L2 ( ,  ) with
r 1
.
f
H2
 f (e i )
L2

ii) If f (e i )   u e i  H 1 / 2 , then its Dirichlet extension into the disc is given by ( z  rei ):

.



1
1
F ( z )   u r e i  ( u z )  ( u  z  )


with
F
2
0

   u

8
2
 f
2
1/ 2
.
Remark 3: The fundamental principle of the Hardy-Littlewood method is the fact, that for N being
an integer it holds
1
e
2iN
0
1 if N  0
.
d  
otherwise
0
They used the principle in the form of the following the

Lemma: Let f ( x)   a n x n with x  1 , then for 0  r  1 it holds
0
r N aN 
1
2
1
 f (re
it
)e iN d .
0
Remark 4: The Voronoi summation formula is related to the Dirichlet divisor problem. The Euler
function  (n) is defined as product of all prime divisors of n ([LaE])
 ( n)
n
:
1
1
(1  )
1 
n m n
p
pn
( m , n ) 1
The error function in the divisor problem is given by ([IvA] chapter 3)
( x) :

'
d (n)  x(log x  2  1) 
n x
1
x
  ( )  O( x  )
4 n x n
with
 ( x) :
where by

'
sin(nx)
1
1
x
 ( x  x)    ( )  2
2
2
2
n
n 1
means that if a or b is an integer than d (a) / 2 or d (b) / 2 .
a  nb
In [NaC] a symmetric and simplified formula of the Voronoi summation formula is given. It requires
that the Fourier cosine transforms are elements of a certain subspace of L2 (0, )
G12 (0, ) : h xh( x)  L2 (0, ) L2 (0, ) .
The definition of G12 (0, ) with xh  L2 is related to the framework of P1. An analogue definition in
the Hilbert space framework of P2 leads to the requirement xh  H 1 / 2 . In this context we note
that

M xh( s)   sM h( s)   x s dh .
0
9
Remark 5: With respect to the Kummer function (P 1) we recall from [LoA] 1.1, [SeA]:
1. The zeros  n of the Kummer function 1 F1 (2ix ) lie in the intervals n  1 / 2, n
2. If 1 F1  L1 ( R) continuous and differentiable and  :1 F1  L2 ( R) , then  is a wavelet.
For the expansion of Kummer functions in terms of Laguerre polynomials and Fourier transforms
we refer to [PiA]. Putting
x
x
and
2
1 3
1
K ( x) :1 F1 ( , ;x 2 )   e t dt
2 2
x0
G ( x) :
e
t 2
dt

the link between the polynomial function z n and the Hermite polynomials H n (z ) is given by ([CaD]):

H n ( x) 
x
 (z  i 2)
n
dG( x)
.

Alternatively we propose a corresponding polynomial system defined by

K n ( x) 
x
 (z  i 2)
n
dK ( x)
.

Remark 6: A function, which is in a sense a generalization of  (s) is the Hurwitz Zeta function,
defined by ([TiE] 2.17:

1
(
n

a) s
n 0
.  ( s, a) : 
for Re( s)  1 , 0  a  1 .
For a  1 resp. a  1 / 2 this reduces to
 (s) , (2 s  1) (s) .
There are also other generalized Zeta function, e.g. Lerch or Epstein or Dedekind Zeta function, as
well as Zeta functions associated with cusp forms ([IvA] 11.8). Related to the Hurwitz Zeta function
is the Dirichlet series ([IvA] 1.8), defined by

 ( n)
n 1
ns
L( s,  ) : 
  (1   ( p) p  s ) 1 for
Re( s)  1
p
where for a fixed q  0  (n) is the arithmetical function known as a character modulo q (for
q  1, L(s,  )   (s) , if (a, q)  1 then  ((a)  0 ). For q  1 it holds
L( s, 1 )   ( s) (1  p  s ) 1   (1  p  s ) 1  (1  p  s ) 1 .
p
p
pq
Thus L( s, 1 ) has a first-order pole at s  1 just like  (s) and it behaves similar to  (s) in many
other ways, while L( s,  ) for   1 is regular for Re( s)  0 .
With respect to the definition of (x) we propose
10
 (n) :
n
n  1/ 2
resp.

L ( s)  
1
 (s)
ns


p
  1 
ps 
p

1
/
2
p 

1
for Re( s)  1 .
The Generalized Riemann Hypothesis (GRH) states that all non-trivial zeros of all Dirichlet Lfunctions have real part equal to ½.
Remark 7: In [NaS] a relationship between the Hilbert space H 1 / 2 , the Teichmüller theory and the
universal period mapping via quantum calculus is given. The Teichmüller spaces are also related
to Riemann surfaces and the geometrization of 3-manifolds.
11
Proof 3+ proposals based on alternative Li-function
We recall from ([BeM]), [EdH]) the following: the density of the primes is the distribution
 ( x)    ( x  p)
.
p
A consequence of the prime number theorem is (e.g. [ViJ])
x
log x
 ( x) 
.
Riemann’s exact formula for  (x) relates to the non-trivial zeros
zn 
1
 it n of
2
the zeta function.
With the notations
M ( x) :
N ( x) : 2e  x / 2
1
1
( x 2  1) x
e
 Im( t n ) x
Re(t n )  0
J ( x) :
cosRe(t n ) x
1
1
M ( x)  N (log x)

log x log x
( x) : xJ ( x) , k ( x) : ( x 1 / k )
.
the Riemann formula is given by

k
k 1
k2
x ( x)  
k ( x)
whereby  k denote the Möbius numbers.
Each term in the sum of N (x) describes an oscillatory contribution to the fluctuations of the
density of primes, with larger Re(t n ) , corresponding to higher “frequencies”. As E (x) decreases
faster than (x) a slower decrease of a infinite series of oscillations N  (x) enables a faster
convergence of corresponding M  (x) , which supports the RH convergence criterion.
If the RH is true, then Im(t n )  0 for all n  N . Then the support of the Fourier transform of N (x)
is discrete, i.e. N (x) has a discrete spectrum, which are the frequencies of a sophisticated
vibration system. Mathematically, these are discrete eigenvalues of a self-adjoint (hermitian)
operator.
Oscillatory integrals are the main subject of Pseudo Differential Operators [PeB]. Singular integral
operators are Pseudo Differential Operators of negative order with corresponding domain. We
claim that the music of the primes is related to the eigenvalues of such a singular integral operator.
Finite element methods provide the appropriate approximation tools to calculate approximation
eigenvalues of corresponding singular integral equations [BrK].
12
Proofs 4+ proposals based on P1/P2 Hilbert space frameworks
The degenerated hypergeometric function allows the definition of an alternative series
representation of the Zeta function, alternatively with respect to Riemann’s Zeta function and to
Polya’s Zeta fake function ([EdH] 1.8, 12.5):

 ( z )  2 (u ) cos(( zu)du .
0
It enables the application of corresponding Polya criterion ([PoG]) for other proofs of the RH.
In the context of the odd (!) function
2
1 3
xe x 1 F1 ( ; , x 2 )
2 2
we refer to the approach in [CaD] (based on [PoG1]) in the framework of Laguerre-Polya class
functions with special functions of genus >1 and its Weierstrass factorization form [CsG]). The
results of [SeA] provides the corresponding nonzero real numbers condition that

1
1


2
 .
The Hilbert transforms of P1 and P2 define appropriate kernels of integral operators with
corresponding domains and ranges defined in appropriate Hilbert spaces. The corresponding
orthogonal systems of the related Hilbert spaces enable the definition of alternative Fourier integral
representations of the Zeta function itself. For the integral operators of P1 and P2 the following
polynomial systems are proposed:
P1: The Hilbert transformed Hermite polynomials
P2: The Lommel polynomials ([DDi], [WaG] 9-7).
With respect to the Bessel function and in relationship to
s
( )
1 3
dx
2
0 x 1 F1 ( 2 ; 2 , x) x  s  1

s/2
and to Fourier transforms of positive definite kernels and the Riemann Zeta function ([PoG1]) we
note ([WaG] 15-53)
Y ( x)
( s)   s 1  x
.
  x e x( J 02 ( x)  Y02 ( x)) d arctan 0
s 1 2 0
J 0 ( x)

13

References
[BaB] B. Bagchi, On Nyman, Beurling and Baez-Duarte’s Hilbert space reformulation of the
Riemann Hypothesis, Indian Statistical Institute, Bangalore Centre, (2005)
[BeB] Berndt B. C., Andrews G. E., Ramanujan´s Notebooks Part I, Springer Verlag, New York,
Berlin, Heidelberg, Tokyo, 1985
[BeM] Berry M. V., Keating J. P., H  xp and the Riemann zeros, in Super symmetry and Trace
Formulae: Chaos and Disorder, Ed. I.V. Lerner, J.P. Keating, D.E. Khmelnitski, Kluver, New York
(1999) pp. 355–367
[BeJ] Bertrand J., Bertrand P., Ovarlez J.-P., The Mellin Transform, in Transforms and Applications
Handbook, ed. Alexander Poularikas, CRC Press, Boca Raton, Florida, 1996
[BrK] Braun K., Interior Error Estimates of the Ritz Method for Pseudo-Differential Equations, Jap.
Journal of Applied Mathematics, 3, 1, 59-72, (1986)
[BrK] Braun K., A distributional Hilbert space framework to leverage the Hardy-Littlewood circle
method, www.riemann-hypothesis.de
[CaD] Cardon D. A., Convolution operators and zeros of entire functions, Proc. Amer. Math. Soc.,
130, 6 (2002) 1725-1734
[CsG] Csordas G., Fourier Transforms of Positive Definite Kernels and the Riemann Zeta-Function
[DDi] D. Dickinson, On Lommel and Bessel polynomials, Proc. Amer. Math. Soc. 5, 946-956 (1954)
[DeJ] Derbyshine J., Prime Obsession, Joseph Henry Press, Washington D.C., 2003
[DoW] Donoghue W. F., Distributions and Fourier Transforms, Academic Press, New York, 1969
[DuR] Duffin R. J., Yukawan potential theory, 1970, Carnegie Mellon University, Department of
mathematical sciences, Paper 161, http://repository.cmu.edu/math
[EdH] Edwards H. M., Riemann´s Zeta Function, Dover Publications, Inc., Mineola, New York,
1974
[EsT] Estermann T., On Goldbach’s problem: Proof that almost all even positive integers are sums
of two primes, Proc. Lond. Soc. 44, 307-314 (1938)
[FlS] Flügge S., Rechenmethoden der Quantentheorie, Springer-Verlag, Berlin, Heidelberg, New
York,1976
[GaD] Gaier D., Vorlesungen über Approximation im Komplexen, Birkhäuser Verlag, Basel, Boston
Stuttgart, 1980
[GrI] Gradshteyn I. S., Ryzhik I. M., Table of Integrals Series and Products, Fourth Edition,
Academic Press, New York, San Francisco, London, 1965
[HaG] Hardy G. H., Littlewood J. E., Some problems of “Partitio Numerorum”; III, On the
expression of a number as a sum of primes, Acta. Math., 44 (1923) 1-70
14
[HaH] Hamburger H., Über einige Beziehungen, die mit der Funktionalgleichung der
Riemannschen Zeta-Funktion äquivalent sind, Math. Ann. 85 (1922) pp. 129-140
[IvA] Ivic A., The Riemann Zeta-Function, Theory and Applications, Dover Publications, Inc.,
Mineola, NewYork, 1985
[LaE] Landau E., Ueber die zahlentheoretische Function  (n) und ihre Beziehung zum
Goldbachschen Satz, Gött. Nachr. 1900, 177-186
[NaC] Nasim C., On the summation formula of Voronoi, Trans. Amer. Math. Soc. 163 (1972) pp.
35-45
[NaS] Nag S., Sullivan D., Teichmüller Theory and the Universal Period Mapping via Quantum
Calculus and the Space on the Circle, Osaka J. Math. 32 (1995) p. 1-34
[PeB] Petersen B. E., Introduction to the Fourier Transform and the Pseudo-Differential Operators,
Pitman Advanced Publishing Program, Boston, London, Melbourne, 1983
[PoG] Polya G., Über die Nullstellen gewisser ganzer Funktionen, Math. Zeit. 2 (1918), 352-383,
also, Collected Papers, Vol II, 166-197
[PiA] Pichler A., Converging Series For The Riemann Zeta Function, arXix:1201.6538v1 [math.NT]
31 Jan 2012
[PoG1] Polya G., Über die algebraisch-funktionentheoretischen Untersuchungen von J. L. W. V.
Jensen, Mathematisk-fysiske Meddelelsev VII 17 (1927), 1-33
[RiB] Riemann, Bernhard. Ueber die Darstellbarkeit einer Function durch eine trigonometrische
Reihe. Göttingen: Dieterich, 1867
[ScW] Schwarz W., Einführung in die Methoden und Ergebnisse der Primzahltheorie, BI,
Hochschultaschenbücher, Mannheim, Wien, Zürich, 1969
[SeA] Sedletskii A. M., Asymptotics of the Zeros of Degenerate Hypergeometric Functions,
Mathematical Notes, Vol. 82, No. 2 (2007) 229-237
[TiE] Titchmarsh E. C., The Theory of the Riemann Zeta-function, Clarendon Press, Oxford, 1951
[ViJ] Vindas J., Estrada R., A quick distributional way to the prime number theorem, Indag.
Mathem., N.S. 20 (1) (2009) 159-165
[ViI] Vinogradov, I. M., Representation of an odd number as the sum of three primes, Dokl. Akad.
Nauk SSSR 15, 291-294 (1937)
[ViI1] Vinogradov, I. M., The Method of Trigonometrical Sums in the Theory of Numbers, Dover
Publications, Inc., Mineola, New York, 2004
[WaG] Watson G. N., A Treatise on the Theory of Bessel Functions, Cambridge University Press,
Cambridge, Second Edition first published 1944, reprinted 1996, 2003, 2004, 2006
[ZyA] Zygmund A., Trigonometric series, Vol. I, Cambridge University Press, 1968
15
Appendix
A.

 1 y
 y e sin(2xy )dy  
2
1 
2
xe x (
2
0

y
 1 y 2
e
cos(2xy )dy 
0
M sin ( s)  ( s) sin(
 1
2
)1 F1 (1 
 3
; ; x 2 )
2 2

, Re(  )  1 , [GrI], 3.952, 7.
, Re(  )  0 , [GrI], 3.952, 8.
1 2 
 1
 ( )1 F1 ( ; ;x 2 )
2
2
2 2

s) , M cos ( s)  ( s) cos( s)
2
2

in the critical stripe ([GrI] 3.952).
,
(1 - s) 
1
   ....
1- s
,
1
1
  ....
s -1 2
 ( s) 
1 

lim  ( s) 

s 1
s  1

lim ( s  1) (s)  1
s 1
 ( x)  x  x  x 
B. The functions
1  sin 2x
1


 ( x)
2 1 2 
2

 H ( x)   log 2 sin(x)  
1
cos 2x

are identical in a weak L#2 (0,1)  sense. It holds
i)
ii)
iii)

 ( s)
s

  x  s  ( x)
0
 ( s)( s) cos(
2

0
 H
2
0

2

for 0  Re( s)  1 , [TiE] 2.1
dx
x

s )   x s  H ( x)
0
2
dx
x
for 0  Re( s)  1
([AmT], [GrI] (4.224).
3
C. [EdH] 12.5, [PoG]: Let f ( x)  
a
1/ a
f (ux) F (u )du define a real self-adjoint operator (where
F (u ) is real and satisfy uF (u)  F (1/ u) ) which has the property that
u F (u ) is
nondecreasing on the interval 1, a  , then its Mellin transform has all its zeros on the critical
line.
D. In a H 1 / 2  sense ([PeB] I, §15) it holds for 0  Re( s)  1

s
 x  H ( x)
0

1  1
dx  1
   x s cos(nx)
 s
x
x
1 n 0
1 n


dy
dy
1 s

s
0 n1 y cos( y) y   (s)0 y cos y y   (s)(s) cos( 2 s)
.
16
The Riemann duality equation is given by


2 ( s) :  ( s) s(1  s)   x s ( xf ( x))d log x   2 (1  s)
0

.
The idea for the proofs P1 and P2 are to replace
P1 :
P2 :
1 s
, M  f (s)  1   2 ( s )  M  f H (s)   2 ( s ) tan( s)
s
f ( x)  f H ( x)
s
2
2
, M  (1  s)  2  ( s  1)  M  H (1  s)   (s)
 ( x)   H ( x)
s 1
2
in the critical stripe.
Note:
1. M h(s)  (1  s)M h(s  1) , M xh(s)  sM h(s) , M ( xh)'(s)  (1  s)M h(s)
2. Hardy ‘s theorem, that the number of zeros of the Zeta function on the critical line is
infinite, provides the “transfer” density argument from weak to strong solution propositions.
3.
M sin ( s)  ( s) sin(
4.
f
2
0
 fH
2
0


2
,
s)
,
1
2

M cos ( s)  ( s) cos( s)
2

2
0
 H
2
0

in the critical stripe.
2
3
5 . [SeA]: All zeros z n of the Kummer function lie in the half-plane Re( z)  1 / 2 and in the
horizontal stripe
(2n  1)  Im( z )  2n .
6.
(1 - s) 
1
   ....
1- s
,
 ( s) 
1
1
  ....
s -1 2
.
E. For the linkage of variational theory to holomorphic function in the distribution sense we
recall from [BPe] chapter I §15 the
Definition: Let z  g z be a function defined on a open subset U  C with values in the
distribution space. Then g z is called a holomorphic in U  C (or g ( z ) : g z is called
holomorphic in U  C in the distribution sense), if for each   C c the function z  ( g s ,  ) is
holomorphic in U  C in the usual sense.
17
F. In the context of proof P2 we recall from [BaB]:
Proposition Let H denote the weighted l 2  space consisting of all sequences a  an n  N  of
complex numbers such that


n 1
n
an
2
 : 11,1,1,......
Let
with c1  n  c2 .
2
2

n
n
 n

 k :   ( ) n  1,2,3,....  H
 k

,
for k  1,2,3,...
and k be the closed linear span of  k . Then the Nyman criterion states that the following
statements are equivalent:
i) The Riemann Hypothesis is true
ii)   k .
G. [EdH] 1.8: The constant not vanishing Fourier term of



1
G( y ) :  f (ny )  1  2 f (ny ) : 1   ( y 2 )  y 1G( y 1 )
requires the definition of an auxiliary function ([EdH] 12.5)



H ( y ) : 2 yG ( y )  y 2 G ( y )  4  (ny ) 2 2 (ny) 2  3 f (ny )  y 1 H ( y 1 ) .
1
This leads to the representation

 ( s)   2(u ) cos(su)du
0
with
2(log x) :
x H ( x) ,
respectively

1
2
 ( s)   a 2 n ( s  ) 2 n
0
with

a 2 n : 4 x
1
3/ 4
1
( log x) 2 n
d x 3 / 2 ( x) dx .
2
(2n)!
dx
x

18

H. Putting

1 3
1 (  x) n
F ( x) :1 F 1( , ; x)  
2 2
n!
1 2n  1
It holds
 2 xF ( x)  e  x
and therefore for 0  Re( s)  1 / 2

( s )   x s e  x
0

(1  s)
dx
s
  * ( s) :  2 x s dF 2
 ( s )
1
x
1

2
s
0
s
2
respectively
( s )
 * ( s) .
s
 *
( s  1) (1  s)  (1  s)
19